MarkItDown vs qmd

Microsoft's MarkItDown is a lightweight Python library that converts virtually any file type — PDFs, Word docs, PowerPoints, Excel spreadsheets, images, audio, HTML, ZIP archives — into clean Markdown optimized for LLM ingestion. It's become one of the most-starred open-source utility tools on GitHub in 2026, surpassing 98,000 stars with a +2,300 gain in a single day. The recent 2026 update added three key features that significantly expand its utility: a Model Context Protocol (MCP) server for direct integration with Claude Desktop and other LLM clients, a plugin-based architecture that lets third-party developers add converters, and fully in-memory processing with no temporary files. The markitdown-ocr plugin extends PDF and Office conversions to extract text from embedded images using LLM vision models. For any developer building RAG pipelines, document QA systems, or LLM-powered data extraction workflows, MarkItDown eliminates the fragmented ecosystem of format-specific parsers. Install only the converters you need, or grab everything with a single pip flag. It's the kind of unsexy infrastructure tool that quietly becomes load-bearing in every serious LLM stack.

qmd is a lightweight local search engine built by Tobi Luetke, CEO of Shopify, for indexing and querying personal knowledge bases, documentation, and meeting notes — entirely offline. It combines three retrieval approaches in a single pipeline: BM25 full-text search for exact keyword matches, vector semantic search via ONNX-based embeddings, and LLM re-ranking using GGUF models through node-llama-cpp. All three stages run locally with no cloud dependency. The tool ships in multiple deployment modes: a CLI for ad-hoc queries, a Node.js library for programmatic use, an HTTP service for local API access, and — most useful for AI workflows — a native MCP server that lets Claude Code, Cursor, and similar editors query your local knowledge base directly during coding sessions. The hybrid retrieval approach means it handles both "find the exact error message from last week's standup notes" and "what was our decision about the auth architecture" equally well. What makes this notable beyond its technical approach is provenance: Luetke shipped it as a personal tool he actually uses, not a startup product. The GitHub history shows active iteration and he's been talking about it on X. It's a credible signal of where pragmatic AI-augmented knowledge management is heading for technical users who prefer local-first tools.